• Added a working example via the zipped 'output_dir' folder for easier testing and understanding.
• Improved GUI styling for better layout, contrast, and usability across sliders and controls.

• Modularized project structure to support easier integration of future models.
• Fixed color issues in generated plots for better visual distinction.
• Temporarily removed parallelization due to issues with multiprocessing and plotting. Sequential execution is now default.

Modifications from v1.0.1:

1. ML-AMPSIT.ipynb has been renamed to ML-AMPSITfast.ipynb to better reflect its faster single-configuration nature compared to ML-AMPSITloop.ipynb, which supports multiple configurations.
2. The GUI in ML-AMPSITfast.ipynb has been enhanced. Users can now configure their setup and use the new RUN/STOP buttons to control execution.
3. The Support Vector Machine's linear kernel has been replaced with a polynomial kernel. Its associated feature importance method is now the Sobol method, consistent with those used for Gaussian Process and Bayesian Ridge.
4. The Gaussian Process kernel has been switched from RBF to Rational Quadratic. A future release will allow the user to select the desired kernel.
5. Added a parallel execution block in ML-AMPSITloop.ipynb using ThreadPoolExecutor to run all the combinations of configurations from loopconfig.json more efficiently.
6. Minor changes to improve the code.
7. A requirements.txt file has been added to facilitate environment setup.
8. Bug Fixes:
   • Removed hardcoded color settings to support plots with any number of parameters.
   • Fixed an issue with the number of simulations (N) slider in the GUI of ML-AMPSIT.ipynb.
   • Plot outputs in ML-AMPSIT.ipynb were erroneously saved to a hardcoded path; they now correctly follow the output_path specified in the configuration file.

Main modifications from v1.0.0:

1. ML-AMPSITloop.ipynb now automatically parallelize the Bayesian search for tuning BayesSearchCV() through the parameter "n_jobs=-1"
2. ML-AMPSIT.ipynb now generates plots with colorblind-friendly colors and improved quality.
3. ML-AMPSITloop.ipynb and ML-AMPSIT.ipynb now automatically normalize sensitivity indices, but the user can find and uncomment the lines "#importance_list.append(importances)" to output non-normalized values (e.g. to visualize the effects of interaction terms over the Sobol total index in the Sobol method)

ML-AMPSIT release version 1.0.0